1

**Work in progress – do not cite without permission**

Estimating values, tradeoffs, and complementarities in ecosystem attributes

Sahan T. M. Dissanayake* and Amy W. Ando

Department of Agricultural and Consumer Economics

University of Illinois at Urbana-Champaign

Abstract: Understanding the value of preserving and restoring ecosystem services is vital for shaping

optimal conservation investments. Recent studies have shown that incorporating public preferences and

economic considerations can lead to a more efficient allocation of resources. We use a choice

experiment survey of Illinois residents and analyze public preferences and willingness to pay (WTP) for

grasslands. We make three contributions to the non-market valuation literature. First, we estimate the

values of multiple facets of grassland ecosystems. Second, we analyze how the quantity of an existing

environmental public good (grasslands) affects the WTP for providing more of that good (restoring new

grasslands). Third, we analyze the public’s willingness to trade off between several common measures of

conservation success (species richness and population density). We find that species richness,

population density, presence of endangered species, presence of wildflowers, and distance from a

respondents home are all significant factors that affect consumers’ WTP for a grassland. This finding

challenges the common practice of using just one of the variables as an indicator of conservation

success. We also find that the presence of nearby grasslands actually increases a respondent’s WTP for

restoring a new grassland. This result is counter to what would be expected from neoclassical economics

and can possibly be explained by endogenous preferences. Further, respondents treat the conservation

success measures (species richness, population density and endangered species) as substitutes for each

other. We finally analyze the impact of including attribute interaction terms on the total willingness to

pay (TWTP) and on the TWTP set of conservation success terms.

Keywords: choice experiment, grassland, conservation, habitat restoration, species richness, population

density, non-market valuation

JEL Codes: Q51, Q57

Acknowledgments: The authors express their gratitude to John B. Braden, Jeff Brawn, Robert Johnston,

James R. Miller, Catalina Londoño, Taro Mieno and participants at the PERE workshop at University of

Illinois for comments and suggestions. Funding for this research was provided by the Research Board at

the University of Illinois and the Robert Ferber Dissertation Award awarded by the Survey Research Lab

at the University of Illinois.

*Corresponding author: Sahan T. M. Dissanayake, Department of Agricultural and Consumer Economics,

University of Illinois, 326 Mumford Hall, MC-710, 1301 West Gregory Drive, Urbana, IL 61801-3605.

Email: sdissan2@illinois.edu

2

1. Introduction

Growing demand for urban space and land for agricultural production has caused much

conversion of many natural habitats, putting pressure on rare and endangered species and

decreasing the flow of ecosystem services. In response, conservation organizations seek to

identify and protect land with high conservation and biodiversity values; this has led to much

research on optimal protected area planning (e.g. Csuti et al. 1997; Ando et al. 1998; Costello

and Polasky 2004; Önal 2004; Önal and Briers 2005; Possingham et al. 2006; Pressey et al.

2007). Most of that research uses production-side factors – the locations of endangered

species, the cost of land, the threat posed to natural areas by development - to guide decisions

about where to locate dedicated natural areas and what features those areas should have.

However, Ando and Shah (2010) show that conservation activity can yield higher social benefits

if decision makers consider the preferences of people when they plan their network of natural

areas. This paper improves understanding of public preferences for conservation by estimating

the values of and substitutability between multiple facets of ecosystems and by analyzing how

the existing quantity of an environmental public good affects WTP for providing more of that

good.

Neoclassical economic consumer theory predicts that marginal willingness to pay for an

increase in a public good will be lower for consumers who already have access to a relatively

large quantity of that good. On the other hand, endogenous preferences can lead to the

opposite effect (Bowles 1998; Zizzo 2003; Gowdy 2004). The theory of endogenous preferences

argues that consumers who are familiar with a good may be willing to pay more than

consumers who have unfamiliar (O’Hara and Stagl 2002). We test this hypothesis by analyzing

3

how the willingness to pay to restore a new grassland is affected by the presence of grassland

areas nearby.

Much of the protected area planning literature uses one indicator, typically species

richness, or a species diversity index such as the Shannon index, as a decision criterion to

identify conservation and restoration projects. At the same time, some studies have shown that

the public has a positive value for species population density. No study has analyzed which

elements of conservation success are valued the most by the public or whether people view

those elements as substitutes or complements. In this paper, we estimate how the marginal

value of any one measure of conservation success - species richness, population density, and

the presence of endangered species - is affected by the levels of the other two.

Though there have been many contingent valuation (CV) studies (and more recently

choice experiment (CE) studies) estimating the values of many kinds of habitat conservation

and restoration, to date no economic valuation study has analyzed the social value of grassland

ecosystems. A massive loss of grassland in North America has stressed wildlife and cut

ecosystem service provision in large swaths of the continent. This problem can be addressed

with grassland restoration activities, but such projects are costly and require difficult and

seemingly arbitrary choices to be made about the exact nature of the grasslands created. The

restoration ecologists who carry out grassland restoration have no guidance from the economic

valuation literature about the preferences people have over the characteristics of restored

grasslands. In this paper we will meet that need for knowledge by using a choice experiment

survey of Illinois residents to analyze willingness to pay (WTP) for grassland habitat restoration.

We find that that species richness, population density, presence of endangered species,

4

presence of wildflowers, and distance from an individual’s home are all significant factors that

affect consumers’ WTP. This challenges the common practice of using just one measure, such as

species richness, as a stand-alone indicator of conservation success. We also find that

respondents with existing grasslands nearby have a higher WTP for restoring a new grassland;

this result is counter to what would be expected from neoclassical economics and can possibly

be explained by endogenous preferences. Finally, the marginal value respondents place on any

one conservation goal (species richness, population density and endangered species) is lower if

the levels of the other two conservation goals are high. This finding implies that respondents

have convex total willingness to pay contours, as opposed to linear or concave. Thus, the

bundle of conservation attributes that maximize TWTP has positive levels for only one of the

conservation success measures (a corner solution where only the number of endangered

species has a positive value). This result changes if physical factors constraint the levels of

conservation success values.

2. Literature Review

Much of the non-market valuation literature on conservation and restoration has

focused on wetland preservation and restoration (Heimlich et al. 1998; Woodward and Wui

2001; Boyer and Polasky 2004), forest preservation and restoration (Adger et al. 1995;

Lehtonen et al. 2003; Baarsma 2003), the protection of individual endangered bird species

(Loomis and Ekstrand 1997; Bowker and Stoll 1988) or recreation and hunting (Hanley et al.

2002; Horne and Petäjistö 2003; Boxall et al. 1996; and Roe et al. 1996). To date no economic

valuation study has analyzed preferences for grassland ecosystems.

Identifying whether existing and new environmental public goods act as substitutes or

5

complements, especially with regard to restoration and conservation of ecosystems and natural

habitats, will enable conservation organizations to better target conservation efforts. Carson et

al. (2001) discuss how public goods will act as substitutes and the WTP will decrease as more of

the public good is provided. This follows from a neoclassical consumer framework that the

demand function is downward sloping. At the same time the presence of an environmental

public good can lead to learning and appreciation such that agents who currently experience

high levels of the public good may have a higher willingness to pay for more of that good

(O’Hara and Stagl 2002). This can be explained using endogenous preferences (Bowles 1998;

Zizzo 2003; Gowdy 2004). A related theory of planned behavior (Ajzen 1991) states that WTP is

expected to increase with a more favorable attitude toward paying for a good (Liebe et al.

2011). Cameron and Englin (1997) analyze the WTP for trout fishing using a CV survey and find

that experience, measured by the number of years in which the respondent has gone fishing,

has a significant positive impact on the WTP. If a favorable attitude towards grasslands can

arise from opportunities to experience existing nearby grasslands, respondents with grasslands

nearby will have a higher WTP to restore a new grassland.

A single measure of conservation success, such as species richness or the number of

endangered species has been used in many ecological and protected areas selection studies

(Csuti et al. 1997; Ando et al. 1998; Haight 2000; Önal, 2004, Cabeza et al. 2004, Kharouba

2010). Studies such as Loomis and Larson (1994) and Fletcher and Korford (2002) also

demonstrate that wildlife population density is an important variable affecting the public’s WTP

for habitats. In terms of maximizing benefits from conservation and restoration it is important

to understand how each of these conservation success measures influences the WTP and how

6

they are related to each other (i.e. do respondents treat the conservation success measure as

substitutes or complements). Christie et al. (2006) study public preferences and WTP for

biodiversity in general and Meyerhoff et al. (2009) find that the species richness is a significant

attributes that determines the WTP for forest conservation. However, neither of the above

studies includes wildlife population density as an attribute, making it difficult to understand the

role that each of these attributes play in determining the WTP for restoration projects.

3. Background: Grasslands Ecosystems

Grasslands are open land areas where grasses and various species of wildflowers are the

main vegetation. In North America there are three main types of grassland ecosystems. The

short-grass ecosystem predominantly occurs on the western and more arid side of the Great

Plains. The mixed-grass ecosystem is located farther to the east. The tall-grass ecosystem

occurs on the eastern side of the Great Plains. Tall grass can grow up to 4-6 feet. Figure 1

presents the distribution of grassland ecosystems in North America (O’Hanlon 2009).

The loss of grassland in North America is attributed to deforestation in the eastern

United States, fragmentation and replacement of prairie vegetation with a modern agricultural

landscape, and large-scale deterioration of western U.S. rangelands (Brennan and Kuvlesky

2005). The loss of grassland ecosystems in most areas of North America has exceeded 80%

since the mid-1800s (Knopf 1994; Noss et al. 1995; Brennan and Kuvlesky 2005). As depicted in

Figure 2.a and Figure 2.b, Illinois has lost 99.9% of its original prairie since the early 1800s, and

currently has 424 state and 24 federally listed threatened and endangered species within its

boundaries (Illinois Wildlife Action Plan 2010).

The loss of grasslands has contributed to a widespread and ongoing decline of bird

7

populations that have affinities for grass-land and grass-shrub habitats (Vickery and Herkert

1999, Askins 2000, Brennan and Kuvlesky 2005). An analysis of continent-wide population

trends on Breeding Bird Survey routes between 1966 and 2002 showed that only 3 of 28

species of grassland specialists increased significantly, while 17 species decreased significantly

(Sauer et al. 2003). During the 25-year period ending in 1984, grassland songbirds (e.g.,

Henslow’s Sparrows, Grasshopper Sparrows, Savannah Sparrows, Bobolinks, Eastern and

Western Meadowlarks, and Dickcissels) in Illinois declined by 75% - 95% (Illinois 1985, Heaton

2000). The Greater Prairie Chicken habitats in Illinois have decreased significantly as evident in

Figure 3 (Ronald 1998). Samson and Knopf (1994) suggest that North America prairies are a

major priority in biodiversity conservation. Fletcher and Koford (2002) perform a comparison of

bird populations in original and restored habitats in Iowa and find that restored grassland

habitats contain bird communities generally similar to those in native prairie. Vickery et al.

(1999) state that given the extent of the decrease in grassland habitat, widespread restoration

of grasslands throughout the US is the most effective approach to restoring bird populations.

In an effort to address these growing concerns, ecologists and conservation biologists

are engaged in restoring grassland habitats to protect endangered flora and fauna, but such

restoration projects are currently informed by knowledge only from the physical, biological, and

ecological sciences (Hatch et al. 1999, Howe and Bron 1999, Fletcher and Korford 2002, Matrin

et al. 2005, Martin and Wilsey 2006). Restoration planners must make choices about exactly

how and where to carry out ecological restoration, and those choices entail physical tradeoffs

between the exact types of restored ecosystems that result, the kinds of animals and plants

that inhabit the restored areas, the variety of species that are supported by the project, the

8

density of wildlife populations that will be present, and the types of management tools used to

maintain these areas. These choices must currently be made in an absence of knowledge about

public preferences regarding the characteristics of grassland restoration projects.

4. Methodology

4.1. Choice Experiment Surveys

CE surveys are being used widely by economists to elicit public preferences for

environmental goods and policies that are typically not related to existing markets (Boxall et al.

1996; Louviere et al. 2000). CE surveys are based on Lancaster’s (1966) consumer theory that

consumers obtain utility from the characteristics of goods rather than the good itself.

Therefore, CEs can be considered the equivalent of hedonic analysis for stated preference

valuation methods. Though CE surveys are more complex to analyze and implement than

contingent valuation studies, they allow the researcher to a detailed understanding of the

respondents’ preferences for the policy or scenario being analyzed. Unlike CV surveys, CE

surveys allow the calculation of part worth utilities for attributes. Hanley et al. (2001) and

Hoyos (2010) provide reviews of the choice experiment methodology.

In a typical CE survey, the respondent repeatedly chooses the best option from several

hypothetical choices that have varying values for important attributes. Choice experiment

surveys require the use of experiment design techniques to identify a combination of attributes

and levels to create the profiles appearing on each survey.

4.2. Survey Instrument

The survey for this research will present respondents with opportunities to express

preferences over pairs of hypothetical restored grasslands that have the following attributes:

9

species richness, wildlife population density, number of endangered species, frequency of

prescribed burning, distance to the site from the respondent’s house, prevalence of

wildflowers, and cost. Some attributes were motivated by our intent to explore preferences

regarding common measures of conservation success. The exact list of grassland attributes was

refined after studying the grassland restoration literature and nonmarket valuation literature.

A CV study on preferences for urban green space in Montpellier, France and a CV study

on preference for protecting or restoring native bird populations in Waikato, New Zealand find

that providing information about the presence of birds significantly effects the WTP. Therefore

we include information about bird species in the survey. A study by Gourlay and Slee (1998) on

public preferences for landscape features find that wildflowers were one of the features most

frequently valued 'highly' or 'very highly'. Since wildflowers are an integral part of the grassland

ecosystems we include the area covered by wildflowers as an attribute. Historically, fire has

been a natural component of grassland ecosystems and many grassland restoration efforts

require management by fire to prevent woody succession and to eliminate invasive species

(Vogl 1979; Schramm 1990; Howe 1995; Copeland et al. 2002). At the same time smoke and ash

from prescribed burns can be hazardous to motorists and become a problem for local residents.

Therefore we include the use of prescribed burns as an attribute in the survey.

Once an initial list of attributes was developed we conducted informal focus groups with

potential survey respondents and discussed the survey with ecologists and land managers at

grasslands. Formal pre-tests of the survey were conducted at the University of Illinois. The final

survey instrument contains background information about grasslands, a description of the

attributes and the levels, 7 sets of binary choice question sets, and a small demographic

10

questionnaire. Appendix A contains an example of one choice question. For each of the binary

choice sets the respondents choose between the two given alternatives and the status quo

option. The choices will contain different features of the restored area and specific values for

these features. The demographic questionnaire has two questions regarding the presence of

nearby grasslands and non-grassland nature areas. The answers to these questions are used to

test whether the presence of nearby grasslands and nature areas has a significant impact on

the WTP to provide a new grassland.

The survey was mailed to a random sample of 2000 addresses in Illinois, stratified

according to population density. The addresses were obtained from the Survey Research Lab at

the University of Illinois. We oversampled from two counties with existing grasslands and two

counties without existing grasslands. We included one dollar bills in half of the surveys to

increase the survey response rate.

4.3. Empirical Design

Given that each choice profile is a binary choice question with a status quo option, a full

factorial survey design would include 36*36*6*6= 19131876 possible profiles. Clearly,

conducting a survey with this many profiles is impractical. Therefore, we follow standard

practice in the choice modeling literature (Adamowicz et al. 1997; Adamowicz et al. 1998;

Louviere et al. 2000) and create an efficient experiment design that will allow both main effects

and interaction effects to be estimated. Given that we are interested in studying the interaction

effects between different indicators of conservation success the design incorporates pairwise

interactions between species richness, population density and number of endangered species.

11

The design for the 7 attributes is presented in Appendix A.1 The design achieves a

99.57% D-efficiency and can be implemented with 54 choice profiles2. The first column in

Appendix A identifies the profile set and the last 7 columns identify the levels of each attribute

that will appear on the survey. We created a block design where the 54 choice sets were

separated into blocks of 6 choice profiles, giving 9 unique surveys with 6 questions each.

Carlsson et al. (2010), show that dropping the first choice question can decrease the error

variance. Therefore, we add an additional choice question before the six choice questions and

drop the first choice question when conducting the analyses to account for possible learning

effects. In order to account for possible ordering effects we reversed the order of the questions

in half the surveys and obtained 18 unique versions of the survey.

4.4. Model and Estimation

CE survey valuation is based on random utility theory (RUM) in which the utility gained

by person q from alternative i in choice situation t is made up of a systematic or deterministic

component (V) and a random, unobservable component (ε) (Rolfe et al 2000; Hensher and

Greene 2001; Hensher et al. 2001).

qit qit qitU V (1)

Following Rolfe et al 2000 and Hensher et al. 2009 the systematic component in (1) can be

separated by the characteristics of the alternative i ( inX ) and the characteristics of the

individual q as below.

1 The experiment design was conducted using the SAS experiment design (Kuhfeld 2005). 2 D-efficiency is the most common criterion for evaluating linear designs. D-efficiency minimizes the generalized variance of the parameter estimates given by D = det *V(X,β)1/k+ where V(X, β) is the variance-covariance matrix and k is the number of parameters (Vermulen et al. 2007; Kuhfeld 2005). Huber and Zwerina (1996) identify four criteria, orthogonality, level balance, minimum overlap and utility balance, which are required for a D-efficient experiment design (Kuhfeld 2005).

12

( , )qit qit qit qitU V X Y (2)

An individual will choose alternative i over alternative j in choice set t if and only if qit qjtU U .

Thus, the probability that person q will choose alternative i over alternative j is given by:

Prob( ; )ij iq iq jq jqP V V j C and j i (3)

where C is the complete set of all possible sets from which the individual can choose. If the

term is assumed to be IIA and Gumbel-distributed the choice probabilities can be analyzed

using a standard multinomial logit model and the probability of choosing alternative i can be

calculated by the following equation where is a scaling parameter (McFadden 1974; Rolfe et

al. 2000; Hensher et al. 2009):

exp( )Prob

exp( )

qit

qit

qjtj C

v

v

. (4)

The standard multinomial logit model generates results in a conditional indirect utility function

of the form,

1 1 2 2 1 2ASC ... ...iq i a b k nV X X Y Y Y (5)

where ASCi is an alternative-specific constant which can capture the influence on choice of

unobserved attributes relative to specific alternatives (Rolfe at al. 2000; Hensher et al. 2009).3

The ’s represent the coefficients on the vector of attributes and individual characteristics. A

willingness-to-pay compensating variation welfare measure can be obtained from the above

estimates as

3 For the empirical specification we do not include an ASC term since the specific alternatives are generic and unlabeled.

13

1

1

cos 0

exp( )

lnexp( )

i

it

i

i

v

WTPv

(6)

where 1

cost is the marginal utility of income (Hanley et al. 2002).4 The part-worth marginal

value of a single attribute can be represented as

cos/k k tWTP . (7)

Though the standard multinomial logit model has been used in many valuation studies

of environmental goods, it assumes that the respondents are homogeneous with regard to

their preferences (the βs are identical for all respondents). This is a strong and often invalid

assumption. Therefore, we use a mixed multinomial logit model5 (Hensher and Greene 2001;

Carlsson et al. 2003) that incorporates heterogeneity of preferences. Assuming a linear utility,

the utility gained by person q from alternative i in choice situation t is given by

qit qi q qit i q qitU X Y (8)

where qitX is a vector of non-stochastic explanatory variables, and

qY is a vector of socio-

economic characteristics. The parameters qi and i represent an intrinsic preference for the

alternative and the heterogeneity of preferences respectively. Following standard practice for

logit models we assume thatqit is independent and identically distributed extreme value type I.

We assume the density of q is given by ( | )f where the true parameter of the

distribution is given by . The conditional choice probability alternative i for individual q in

4 The ’s represent marginal utilities (k=U/Zk) 5 Also referred to as mixed logit, hybrid logit and random parameter logit, random coefficient logit model

14

choice situation t is logit6 and given by

exp( )( )

exp( )

qi q qit i q

q q

t qj q qjt j q

j J

X YL

X Y

. (9)

The unconditional choice probability for individual q is given by,

( ) ( ) ( | )q qP L f d . (10)

The above form allows for the utility coefficients to vary among individuals while

remaining constant among the choice situations for each individual (Carlsson et al. 2003). There

is no closed form for the above integral therefore qP needs to be simulated. The unconditional

choice probability can be simulated by drawing R drawings of , r , from ( | )f 7 and then

averaging the results to get

1( ) ( )q q r

r R

P LR

. (11)

The interpretation of the coefficient values for the above mixed multinomial model is

complicated. Therefore following Carlsson et al. (2003) we calculate the marginal rates of

substitution between the attributes using the coefficient for cost as numeraire and we can

interpret the ratios as average marginal WTP for a change in each attribute.

4.5. Econometric Specification

We use three econometric specifications to test for robustness of the results and to

incorporate individual heterogeneity.

The conditional logit is:

6 The remaining error term is iid extreme value. 7 Typically ( | )f is assumed to be either normal or log-normal but it needs to be noted that the results are

sensitive to the choice of the distribution.

𝑉𝑛𝑖 = 𝛽1𝑋𝑟𝑖𝑐ℎ𝑛𝑒𝑠𝑠 + 𝛽2𝑋𝑑𝑒𝑛𝑠𝑖𝑡𝑦 + 𝛽3𝑋𝑒𝑛𝑑𝑎𝑛𝑔𝑒𝑟𝑒𝑑 + 𝛽4𝑋𝑤𝑖𝑙𝑑𝑓𝑙𝑜𝑤𝑒𝑟𝑠+ 𝛽5𝑋𝑏𝑢𝑟𝑛𝑖𝑛𝑔 + 𝛽6𝑋𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 + 𝛽7𝑋𝑐𝑜𝑠𝑡 + 𝜀𝑛𝑖

15

(12) The mixed multinomial logit is:

(13)

The mixed multinomial logit with interaction terms is:

(14)

This most complex specification (14) includes three variables that are interactions

between the conservation attributes. A significant and positive coefficient on the interaction

term implies that the respondent has higher marginal utility for increases in one conservation

success measure when the levels of the other conservation success terms are high. A significant

and negative coefficient on the interaction terms implies the opposite.

This specification also interacts the cost attribute with person-specific dummy variables

that indicate the presence of grasslands and the presence of non-grassland natural areas

nearby. They allow us to analyze the impact of existing natural areas on the WTP for a new

hypothetical grassland. If the coefficient is positive (negative) and significant this implies that

respondents who have a nearby natural area are willing to pay more (less) to restore a new

grassland.

The conditional logit model was estimated using the built in function within STATA. The

mixed multinomial logit and the mixed multinomial logit with interaction terms were estimated

using the user written STATA routine by Hole (Hole, 2011).

𝑉𝑛𝑖 = 𝛽1𝑛𝑋𝑟𝑖𝑐ℎ𝑛𝑒𝑠𝑠 + 𝛽2𝑛𝑋𝑑𝑒𝑛𝑠𝑖𝑡𝑦 + 𝛽3𝑛𝑋𝑒𝑛𝑑𝑎𝑛𝑔𝑒𝑟𝑒𝑑 + 𝛽4𝑛𝑋𝑤𝑖𝑙𝑑𝑓𝑙𝑜𝑤𝑒𝑟𝑠+ 𝛽5𝑛𝑋𝑏𝑢𝑟𝑛𝑖𝑛𝑔 + 𝛽6𝑛𝑋𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 + 𝛽7𝑛𝑋𝑐𝑜𝑠𝑡 + 𝛽8𝑛𝑋𝑟𝑖𝑐ℎ𝑛𝑒𝑠𝑠 ∗ 𝑋𝑑𝑒𝑛𝑠𝑖𝑡𝑦+ 𝛽9𝑛𝑋𝑑𝑒𝑛𝑠𝑖𝑡𝑦 ∗ 𝑋𝑒𝑛𝑑𝑎𝑛𝑔𝑒𝑟𝑒𝑑 + 𝛽10𝑛𝑋𝑒𝑛𝑑𝑎𝑛𝑔𝑒𝑟𝑒𝑑 ∗ 𝑋𝑟𝑖𝑐ℎ𝑛𝑒𝑠𝑠+ 𝛽11𝑛𝑋 𝑐𝑜𝑠𝑡 ∗ 𝑌 𝑔𝑟𝑎𝑠𝑠𝑙𝑎𝑛𝑑 𝑛𝑒𝑎𝑟? + 𝛽12𝑛𝑋 𝑐𝑜𝑠𝑡 ∗ 𝑌 𝑛𝑎𝑡𝑢𝑟𝑒 𝑟𝑒𝑠𝑒𝑟𝑣𝑒 𝑛𝑒𝑎𝑟?+ 𝜀𝑛𝑖

𝑉𝑛𝑖 = 𝛽1𝑛𝑋𝑟𝑖𝑐ℎ𝑛𝑒𝑠𝑠 + 𝛽2𝑛𝑋𝑑𝑒𝑛𝑠𝑖𝑡𝑦 + 𝛽3𝑛𝑋𝑒𝑛𝑑𝑎𝑛𝑔𝑒𝑟𝑒𝑑 + 𝛽4𝑛𝑋𝑤𝑖𝑙𝑑𝑓𝑙𝑜𝑤𝑒𝑟𝑠+ 𝛽5𝑛𝑋𝑏𝑢𝑟𝑛𝑖𝑛𝑔 + 𝛽6𝑛𝑋𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 + 𝛽7𝑛𝑋𝑐𝑜𝑠𝑡 + 𝜀𝑛𝑖

16

5. Results and Discussion

Out of the 2000 surveys that were mailed out, 300 surveys were returned out of which

263 were complete yielding 1578 choice question observations. Each of the 18 different survey

versions were returned at least 10 times.8 This ensures that each of the 54 choice profiles were

represented in the final analysis. Of the 300 surveys that were returned, 187 were surveys that

included the dollar bill. Therefore, including the dollar bill increased the response rate by 65%.

5.1 Testing for Preference Stability.

There is an ongoing discussion in the non-market valuation literature regarding

preference stability in choice experiment surveys with repeated choices. Though many studies

(Carlsson and Martinsson 2001; Johnson and Bingham 2001; Hanley et al. 2002; Homes and

Boyle 2005; Clark and Friesen 2008; Ladenburg and Olsen 2008; Savage and Waldman 2008;

Bateman et al. 2008; Bush et al. 2009; Brouwer et al. 2010; Carlsson et al. 2010, Day and Prades

2010) have analyzed ordering and learning effects in choice experiment surveys, there is no

clear consensus on the presence of ordering and learning effects. We contribute to this ongoing

discussion by using a novel experiment design to test for ordering and learning effects in choice

experiment surveys.

The experimental design for the choice experiment survey required 9 unique surveys

each with 6 choice questions. We created the experimental framework for testing learning and

ordering effects by first creating a second set of surveys by reversing the order of the 6 choice

questions and second by repeating the first choice question at the end. This gives us 18 unique

8 On average each survey versions was returned 16.6 times with a standard deviation of 0.88 a minimum of 10 and a maximum of 24.

17

surveys with 7 choice questions in each survey.

The results for testing for learning are shown in Table 2.a. The results corresponding to

dropping the first choice are on the left and the results corresponding to dropping the last

choice are on the right. The significance of two of the interactions terms changes between the

two sets of results, which implies that there are significant learning effects. In contradiction to

Carlsson et al. 2010, we find that dropping the first choice question does not significantly

influence the error variance of the estimates..

The results for testing for ordering effects are shown in Table 2.b. The first sets of

results correspond to the initial 9 versions of the survey. The second sets of results correspond

to the second 9 versions of the survey where the order of the choice questions were reversed.

The columns again refer to dropping the first choice and the last choice respectively. There are

noticeable ordering effects. For the first 9 versions of the survey, two of the interaction terms

are not significant whereas when the order of the choice questions is reversed these variables

become significant. We believe that this indicates there are ordering effects but they are

limited to the interactions terms. We drop the first choice question to account for learning

effects and use the pooled data from all 18 versions for the remainder of the analysis to

overcome ordering effects.

5.1 Main Results.

The results for the main-effects regressions (conditional logit and mixed logit) models

are presented in Table 3. These specifications do not include interaction terms. The last column

of Table 3 indicates that individual heterogeneity is significant for many attributes and should

be taken into consideration. However, the parameter estimates are qualitatively similar across

18

the two models. The three conservation attributes and wildflowers all have positive and

significant coefficients, while distance and cost are negative and significant.

For each set of results we calculate the marginal willingness to pay (MWTP) for each

attribute by dividing the coefficient for each attribute by the coefficient for cost as

𝑀𝑊𝑇𝑃 =

(15)

The resulting MWTP values are shown in Table 4. Though the coefficient values for the

conditional logit and the mixed logit models vary in magnitude, the MWTP values for each

attribute is similar for both models. The coefficient estimates should not be compared to each

other directly since the units for each attribute differ. All three of the conservation success

measures (species richness, population density and endangered species) have significant per

household values. A typical person is willing to pay $1.13 each year to have an additional bird

species present in the grassland, and the value of an endangered species is a much higher

$9.09, while increasing the population density of birds in a grassland by 1 additional bird per

acre is worth $1.60. This latter result reinforces the findings by Loomis and Larson (1994) and

Fletcher and Korford (2002) that wildlife population density is an important variable affecting

the public’s WTP for restoring habitats.

The results for the mixed logit model with interaction terms are presented in Table 5.

The coefficient for the interaction of the cost and the grassland near variable is negative. This

implies that respondents who live near existing grassland areas have a higher MWTP for each of

the attributes. This result contradicts what would be predicted by standard neoclassical

consumer economics. This finding could be evidence of endogenous preferences - individuals

who consume and experience a good can have a higher WTP for the good than individuals who

19

have not experienced a good. It could alternatively be argued that this result is caused by

locational sorting wherein respondents who have an inherent preference for grasslands choose

to live close to them. We note that people with high values for grasslands may also have

relatively high values for other natural areas, but the interaction effect for non-grassland

natural areas being nearby is not significant in the regression; this might imply that the positive

coefficient on the interaction of cost with the grassland near dummy is more likely caused by

endogenous preferences than by sorting.

The two-way interaction terms between species richness, population density and

endangered species are all significant and negative; the marginal value of one conservation

feature is lower when the levels of the other feature is high. Figure 4 shows the TWTP as a

function of species richness for different levels of population density. The TWTP increases as

the value of species richness increases. The three lines in Figure 4 correspond to different

levels of population density. As population density increases the TWTP at each level of species

richness increases. When the interaction terms are set to zero (Figure 4.a) the increase in TWTP

caused by higher population density is the same at every species richness level (the lines are

parallel). When the interaction terms are included (Figure 4.b), the slope of the TWTP-species

richness line decreases as the level of population density increases. This illustrates the

relationship between preferences over any two conservation goals; here an increase in species

richness has a smaller impact on TWTP at high levels of population density than at low levels of

population density.

Further, the significant interaction terms implies that the total WTP (TWTP) curves are

non-linear as depicted in Figure 5, which depicts the TWTP contour in species richness and

20

population density space (similar to a utility function in two good space). Figure 5.a contains a

TWTP contour for a TWTP of $80. This contour shows the combination of species richness and

population density that yield a TWTP of $80.When the interaction terms are ignored the TWTP

contour is linear, indicating a fixed marginal rate of substitution. When the interaction terms

are included the TWTP contour is concave, indicative an increasing marginal rate of

substitution. Figure 5.b shows the substitution between species richness and population

density for different levels of TWTP and number of endangered species. The TWTP contour for

$70 lies below the TWTP for $80. As the value of endangered species increases, the TWTP

contour shifts inwards since a smaller amounts of species richness and population density are

required to reach the $70 TWTP contour.

Next we characterize the bundle of conservation success attributes that will provide the

largest TWTP while holding other attributes of a grassland constant. We solve a simple

constrained maximization problem where the TWTP is maximized as a function of the

conservation success variables. The results are presented in Table 7. The first column indicates

whether physical constraints are present; the first sets of results are unconstrained while the

second sets of results assume physical limits on the levels of some attributes. The second

column indicates the budget constraint and the third column indicates the stylized costs. We

assume that each of the conservation success attributes can be produced independently and

that the costs are given per unit of each each attribute. We solve the problem for a range of

total cost values to show how the result changes with the cost. Column four indicates whether

the results include the interaction terms. Columns five through seven report the resulting

optimal values of the conservation success variables and column eight contains the

21

corresponding TWTP amount.

The first sets of results correspond to a scenario without physical constraints. When the

costs are all $1 (the cost ratio is1: 1: 1), for both the scenarios with and without interaction

terms, the result is a corner solution where only the endangered species variable has a positive

value. This is to be expected given the concave TWTP curves and the fact that endangered

species has the highest marginal value. Given that it is relatively difficult to manage a grassland

to attract endangered species, we increase the relative cost of endangered species. When the

cost of endangered species in increased to $10 (cost ratio of 1:1: 10), the solution changes so

that only the population density variable has a positive value. Again this result makes sense

since population density has the second largest marginal value. These corner solutions are to be

expected given the nature of the indifference curves depicted in Figure 4. Given the slope of the

cost function the unbounded utility maximizing bundle will consist of just one attribute. The

above results are not affected by the interaction terms.

Next we present the TWTP-maximizing bundle when physical constraints are imposed

on the levels of conservation goals that can be achieved. The results show that the TWTP-

maximizing bundle is one with high values for the attributes that have a higher marginal

contribution to the overall TWTP. For example, when the budget is unconstrained the scenario

without interaction terms selects the maximum possible values for each of the three

conservation success attributes. When the interaction terms are included, only the population

density variable and the endangered species variable have positive values. This result makes

sense since the interaction terms are negative and if the species richness variable had a positive

value the net effect of its presence would be a decrease in TWTP due to the interaction terms.

22

When the budget is constrained, for the scenario without interaction terms, the TWTP

maximizing solution is the solution to the knapsack problem.9 For the scenario with interaction

terms, the conservation success variable with the lowest contribution, species richness, has

zero value.10These results highlight that though all three conservation success variables are

significant, from the point of view of maximizing TWTP, only the variables that have the highest

marginal contribution to TWTP will have a positive value. These results are a simplification since

there are physical relationships between each of these variables (it is not possible to have a

grassland with endangered species but without any species richness) but the results inform

policy makers on what characteristics of grasslands are most important with regard to

maximizing public support for restoration activities.

Finally, we calculate the total WTP (TWTP) for a hypothetical grassland with realistic

attribute values. Due to the various interaction terms, the result is best represented as the

table shown in Table 6. We estimate the TWTP for a 100 acre hypothetical grassland with 30

different bird species, 15 individual birds per acres, 6 endangered species, 60% wildflower

coverage, and controlled burning once every year and when no non-grassland nature area is

nearby. The TWTP ranges between $60 and $109 per household per year. The results indicate

that being near an existing grassland increases the TWTP for an additional grassland by as much

as 43% (when the new grassland is 10 miles away). Further, as the distance to the restored

grassland increases from 10 miles to 100 miles the TWPT decreases by as much as 28%.

9 Obtain as much as allowed from the attribute that has a highest marginal contribution to the objective function, then as much as allowed from the attribute with the second highest marginal contribution and so on. 10 If the species richness value also had a positive amount the net effect will be a decrease in TWTP due to the interaction terms

23

6. Conclusion

We analyze the structure of public willingness to pay for different attributes of grassland

ecosystems using a choice experiment survey. This work yields several findings that have broad

implications for conservation planning and environmental valuation. First, we find that several

features of an ecosystem that are used as measures of conservation success - species richness,

population density, and presence of endangered species - have large positive marginal values.

Much of the work on optimal protected-area planning and design uses a single measure of

conservation success as the objective to be maximized. Our results imply that when there are

physical tradeoffs between conservation outcomes (e.g. one can increase the population of a

single species such as pheasant, but in doing so one might lower species richness) planners

should be careful to consider all conservation success measures in order to maximize the social

welfare obtained from conservation and restoration efforts.

Second, in an effort to analyze the structure of the preferences for the conservation

success attributes in more detail we use a specification that contains pairwise interactions of

the conservation success terms. We find that the values people place on any one conservation

outcome is lower when the levels of other conservation outcomes are high; in other words,

people seem to view these feature as substitutes rather than complements. This means, for

example, that the value to society of a project that maximizes species richness will vary across

sites that have different levels of wildlife population density and numbers of endangered

species.

Given that restoration ecologists are able to determine the levels of species richness,

population density and the presence of endangered species when undertaking conservation

24

efforts, our results emphasize the importance of considering the levels of all these attributes

when conducting restoration efforts, optimal protected area planning models, and cost benefit

analysis for conservation and restoration of ecosystems.

We also show that as a result of the concave TWTP contours, the TWTP maximizing

grassland only has positive values for one of the conservation success terms (there is a corner

solution). If the signs on the interaction terms were reversed, i.e. the willingness to pay for a

given attribute increased with the values of the other attributes, the indifferences curves would

be convex and this would have resulted in interior solutions with positive values for multiple

conservation success attributes.

Third, we find that respondents who live near existing grassland areas have a higher

MWTP for restoring additional grasslands. This result contradicts what would be predicted by

standard neoclassical economics- the marginal value of a good will decline with its total

quantity. This would mean that the MWTP for grasslands, wetlands, and any number of

environmental public goods would decline with the quantity of those goods that are provided.

Our result may reflect the existence of endogenous preferences - individuals who consume and

experience a good learn to appreciate and enjoy it and can therefore can have a higher WTP

than individuals who have not experienced the good.

We recognize that this finding could be caused by locational sorting; however, as

discussed earlier, the fact that the WTP for an additional grassland is not correlated with the

presence of nearby non-grassland natural areas leads us to believe our result is evidence of

endogenous preferences. We will control for possible endogeneity of proximity to grassland in

future versions of this work. If the result is robust, it has implications for conservation planning

25

in terms of locating new conservation areas; for example the welfare maximizing conservation

strategy may be to have similar ecosystem types partially clustered in the landscape.

Finally, this study is the first to generate value estimates for the WTP to conserve and

restore grasslands, an ecosystem type that is disappearing throughout North America. This

study provides valuable information to conservation planners and ecologists engaged in

restoring and conserving ecosystems regarding the values placed on grasslands by the public.

The results allow policy makers to calculate the total willingness to pay for a grassland with

varied characteristics. For the grassland described in the results section, the annual value per

household ranges between $60 and $109.11 This information is especially important in places

like Illinois where some lands could be potentially be restored as wetland, tallgrass prairie or

forest with different restoration and management techniques. The annual TWTP value per

household that we obtain above is similar to WTP figures obtained for other ecosystem types.12

These values are large enough to exceed restoration costs in many places. To illustrate,

we calculate the TWTP to restore a grassland by each county in Illinois.13 The results are shown

in Figure 6. Champaign County, for example, has a TWTP of $5,344,880 for all households in the

county. The costs for restoring a prairie range from $350 per acre to over $8,500 per acre

(excluding cost of land and maintenance) depending on the existing land conditions and the

11 For a 100 acre hypothetical grassland with 30 different bird species, 15 individual birds per acres, 6 endangered species, 60% wildflower coverage, and controlled burning once every year. 12 Boyer and Polasky (2004) give examples of stated preference surveys that yield WTP for wetlands in the range of $15 (1987$) - $87 (1998$) per hectare per year. Brander et al (2006) conduct a comprehensive summary of stated preference studies on wetlands and find the median willingness to pay is approximately 200 1995 $ per hectare per year. Heimlich et al. (1998) find empirical estimates of the WTP for wetlands that range between $0.02 to $8,924 per hectare. Barrio and Loureiro (2010) conduct a meta-analysis of CV studies of forests and find values that range between $0.75 (ppp 2008$) - $490 (ppp 2008$). 13 We assumed the per-household willingness to pay is $71 (the average between the highest and lowest values from Table 5). We also assumed that the population is located at the center of the county.

26

number of prairie plants and wildflowers being planted (Prairie 2010). Therefore restoring a 100

acre prairie will cost at most $850,000 plus the cost of the land. Assuming the grassland is being

restored from land purchased at the average farmland price in 2010 of $4820 per acre14, the

total cost of restoring a 100 acres prairie is 1,332,000. This illustration implies that restoring a

prairie passes a cost-benefit analysis in Champaign County as well as nearly one third of the

counties in Illinois.15 Our results indicate that the social value of a grassland is higher when

people are close to it; thus, the cost-benefit calculation for restored grasslands will be more

favorable if managers incorporate the presence of individuals when identifying the locations for

conservation and restoration efforts (Ando and Shah 2010).

The results we present allow conservation organizers and land use planners to

effectively conduct a cost benefit analysis of restoring grasslands, and improve restoration

planning decisions about the attributes of restored grasslands. The findings also raise

provocative questions about the standard economic assumption that marginal value of

environmental goods diminishes with total quantity; those questions should be further

explored in future research.

14 The average farmland prices were obtained from the University of Ilinois FarmDoc website (http://www.farmdoc.illinois.edu/manage/newsletters/fefo10_17/fefo10_17.html). We use the average land price for Illinois to illustrate how the data can be used. Real land prices are homogeneous across counties; the heterogeneity should be taken into consideration for a more detailed analysis. 15 Of the 102 counties, 31 had a county TWTP greater than $1,332,000.

27

7. Tables and Figures

Figure 1: Grasslands in North America

Source: (Nature Conservancy, 2008)

28

Figure 2: Grasslands in Illinois

Panel 2.a: Map of Historic Prairies in Illinois from 1820 (Anderson 1970)

Panel 2.b Map of Rural Grassland in Illinois from 2000(Created from INGDCH Land Cover data)

29

Figure 3: Distributions of the greater prairie chicken in Illinois (Anderson 1970)

30

0 5 10 15 20 25 3020

40

60

80

100

120

140

160

Species Richness

TW

TP

Population Density = 0

Population Density = 5

Population Density = 10

Population Density = 15

Figure 4: Species Richness vs TWTP as Density Changes

a. With interaction terms set to zero

b. With positive interaction terms

0 5 10 15 20 25 3020

40

60

80

100

120

140

160

Species Richness

TW

TP

Population Density = 0

Population Density = 5

Population Density = 10

Population Density = 15

0 5 10 15 20 25 3020

30

40

50

60

70

80

90

100

Species Richness

TW

TP

Density = 0

Density = 5

Density = 10

Density = 15

0 5 10 15 20 25 3020

30

40

50

60

70

80

90

100

Species Richness

TW

TP

Density = 0

Density = 5

Density = 10

Density = 15

31

Figure 5.b: TWTP with and without interaction terms

Figure 5.a: TWTP Curves over changing attribute values

0 5 10 15 20 250

2

4

6

8

10

12

14

Popula

tion D

ensity

Species Richness

twtp = 80 w int

twtp = 80 w/o int

0 2 4 6 8 10 12 14 16 18 200

5

10

15

Species Richness

Popula

tion D

ensity

TWTP = 80 endangered = 2

TWTP = 70 endangered = 2

TWTP = 70 endangered = 3

32

Figure 6: TWTP by Illinois County

County TWTP

$166,921 – 1,016,010

$1,016,010 – 2,883,310

$2,883,310 – 6,596,468

$6,596,468 – 12,462, 204

$12,462, 204 – 23, 828,091

$23, 828,091 – 148,824,591

33

Table 1: Attributes and levels for survey instrument

Attribute Description Levels

Number of

Bird Species

The number of different bird species in the restored area. A high number means you are more likely to see many different kinds of birds in the restored area.

30 different species

20 different species

10 different species

Density of

Birds

The number of individual birds (from all species) within an acre. A high number means you are more likely to see a large number of individual birds in the restored areas. They may be all the same type, or they may be several different types.

15 individuals per acre

10 individuals per acre

5 individuals per acre

Number of

endangered

species

The number of different endangered or threatened bird species that will live in the restored area.

6 endangered or threatened species

3 endangered or threatened species 0 endangered or threatened species

Amount of

wildflowers

The percentage of restored land area that will be covered by wildflowers. A higher percentage means you are more likely to see more wildflowers in the restored area.

60% covered in wildflowers

40% covered in wildflowers

20% covered in wildflowers

Use of

prescribed

burning

The possible use of prescribed burns to manage the grassland.

No prescribed burning

Prescribed burning once every other year.

Prescribed burning once every year

Distance to

restored

area

The distance to the restored area from your home. This feature ranges from 10 miles (between 8 to 12 minutes) to 100 miles (between 1 1/2 to 2 hours)

10 miles

50 miles

100 miles

Annual cost

to your

household

The fee that your household will have to pay every year to restore and maintain the grassland.

This value will range from $0 to $100

34

Table 2.a: Testing for Learning

Variable Drop First Choice Drop Last Choice

Coefficient SE Coefficient SE

Main Effects Species richness 0.140 *** 0.024

0.133 *** 0.023

Population density 0.308 *** 0.044

0.346 *** 0.048

Endangered Species 0.568 *** 0.115

0.696 *** 0.135

Wildflowers 0.017 *** 0.005

0.020 *** 0.005

Prescribed burning -0.067 0.093

-0.172 0.106

Distance -0.013 *** 0.002

-0.016 *** 0.003

Cost -0.062 *** 0.008

-0.052 *** 0.006

Richness X Density -0.011 *** 0.002

-0.011 *** 0.002

Density X Endangered -0.014 0.009

-0.026 *** 0.009

Endangered X richness -0.008 * 0.004

-0.007 0.004

Number of O bservations 4734

4722

LR chi2(7) 791.64

790.7 Prob > chi2 0.00 0.00

***significant at 1%, **significant at 5%, *significant at 10%

35

Table 2.b: Testing for Ordering

Original Ordering Variable Drop First Choice Drop Last Choice

Coefficient SE Coefficient SE

Main Effects Species richness 0.173 *** 0.042

0.175 *** 0.039

Population density 0.328 *** 0.068

0.331 *** 0.077

Endangered Species 0.494 *** 0.191

0.393 ** 0.177

Wildflowers 0.018 ** 0.008

0.018 *** 0.007

Prescribed burning 0.012 0.142

0.015 0.146

Distance -0.017 *** 0.006

-0.015 *** 0.004

Cost -0.046 *** 0.009

-0.052 *** 0.010

Richness X Density -0.013 *** 0.004

-0.014 *** 0.004

Density X Endangered -0.006 0.014

0.006 0.014

Endangered X richness -0.011 0.008

-0.007 0.007

Number of O bservations 2358

2352

LR chi2(7) 428.9

453.12 Prob > chi2 0.00 0.00

***significant at 1%, **significant at 5%, *significant at 10%

Reversed Ordering Variable Drop First Choice Drop Last Choice

Coefficient SE Coefficient SE

Main Effects Species richness 0.141 *** 0.033

0.164 *** 0.035

Population density 0.368 *** 0.068

0.355 *** 0.063

Endangered Species 0.880 *** 0.197

0.858 *** 0.178

Wildflowers 0.018 *** 0.007

0.015 * 0.008

Prescribed burning -0.214 0.149

-0.071 0.138

Distance -0.016 *** 0.004

-0.016 *** 0.004

Cost -0.069 *** 0.009

-0.059 *** 0.008

Richness X Density -0.011 *** 0.003

-0.012 *** 0.003

Density X Endangered -0.034 *** 0.013

-0.033 *** 0.012

Endangered X richness -0.010 0.007

-0.011 * 0.007

Number of O bservations 2376

2370

LR chi2(7) 367.7

367.23 Prob > chi2 0.00 0.00

***significant at 1%, **significant at 5%, *significant at 10%

36

Table 3: Regression Results for the Conditional Logit and Mixed Logit Models

Variable Conditional Logit Mixed Logit

Coefficient SE Coefficient SE SD^

Species richness 0.017 *** 0.004 0.029 *** 0.010 Significant

Population density 0.024 *** 0.008 0.092 *** 0.018

Endangered Species 0.135 *** 0.014 0.321 *** 0.043 Significant

Wildflowers 0.013 *** 0.002 0.032 *** 0.005 Significant

Prescribed burning -0.016 0.042 0.121 0.099 Significant

Distance -0.005 *** 0.001 -0.011 *** 0.003 Significant

Cost -0.015 *** 0.001 -0.042 *** 0.005 Significant

Number of Observations 4734

4734

Log Likelihood -1534.26

-1169.70

LR chi2(7) 398.70

729.13

Prob > chi2 0.00 0.00

***significant at 1%, **significant at 5%, *significant at 10% ^Significance of standard deviations at 10% or less when incorporating individual heterogeneity

Table 4: Marginal Willingness to Pay Estimates

Attribute Clogit Mixlogit

Species Richness $1.13 $0.71

Bird Density $1.60 $2.22

Endangered Birds $9.09 $7.73

Wildflowers $0.86 $0.77

Burning -$1.08 $2.92

Distance -$0.31 -$0.25

37

Table 5: Results for the Mixed Logit Model with Interaction Terms

Variable Mixed Logit with Interaction Terms

Coefficient Standard

Errors SD^

Main Effects

Species richness 0.155 *** 0.028 Significant

Population density 0.337 *** 0.052 Significant

Endangered Species 0.692 *** 0.140 Significant

Wildflowers 0.020 *** 0.006 Significant

Prescribed burning -0.025 0.110 Significant

Distance -0.015 *** 0.003 Significant

Cost -0.084 *** 0.015 Significant

Conservation Success Interaction Terms

Richness X Density -0.012 *** 0.003 Significant

Density X Endangered -0.017 * 0.010

Endangered X richness -0.009 * 0.005 Significant

Complementarity Interaction Terms

Grassland Near X Cost 0.025 ** 0.011

Nature Near X Cost 0.009 0.014 Significant

Number of Observations 4734

Log Likelihood -1112.74

LR chi2(7) 811.34

Prob > chi2 0.00

***significant at 1%, **significant at 5%, *significant at 10% ^Significance of standard deviations at 10% or less when incorporating individual heterogeneity

38

Table 6: TWTP for a Hypothetical Grassland

Distance Grassland Near

0 1

10 $66 $93

(47-85) (47-140)

100 $49 $70

(33 - 65) (33 - 107) Note: The 95% confidence interval for each estimate is given within the parentheses.

39

Table 7: Constrained Maximum TWTP

16 0 ≤ Species richness ≤ 30, 0 ≤ Population density ≤ 15, 0 ≤ Endangered Species ≤ 6,

Physical

Constraints

Budget

Constraint

Cost

Ratio

Interaction

Terms

Richness Density Endangere

d

TWTP

None $100 total Equal (1:1:1) No 0 0 100 $1172.3

Yes 0 0 100 $1172.3

$100 total 1:1:10 No 0 100 $569.9

Yes 0 100 0 $569.9

Application Unconstrained Equal (1:1:1) No 30 15 6 $234.2

bounds16 Yes 0 15 6 $130.12

$30 total Equal (1:1:1) No 9 15 6 $179.22

Yes 0 15 6 $130.12

$15 total Equal (1:1:1) No 0 9 6 $106.12

Yes 0 9 6 $121.45

40

8. Appendix

8.1. Appendix A: Survey Design for 7 attributes

Set x1 x2 x3 x4 x5 x6 x7 Set x1 x2 x3 x4 x5 x6 x7 Set x1 x2 x3 x4 x5 x6 x7

1 2 1 2 3 1 2 6 1 2 1 2 3 1 4

2 3 1 2 2 3 1 5 2 2 1 1 2 3 2

3 2 3 2 1 2 3 5 1 2 1 2 3 1 4

4 3 3 1 2 1 3 2 2 2 3 1 3 2 4

5 1 3 1 1 2 3 3 3 2 2 2 1 2 6

6 1 1 3 3 3 3 6 3 3 1 2 2 2 5

7 2 2 2 3 2 1 1 1 3 3 2 3 3 2

8 3 2 1 3 1 1 1 1 1 3 2 2 2 4

9 3 3 2 1 1 1 1 1 1 3 3 3 3 6

10 3 1 2 2 3 1 5 2 2 1 1 2 3 2

11 3 1 2 3 2 2 2 2 3 1 1 3 1 6

12 1 3 2 3 2 2 6 3 1 3 1 1 3 3

13 3 2 3 3 3 3 5 1 3 2 2 1 1 4

14 1 1 2 1 3 3 1 2 3 3 3 2 2 3

15 2 1 1 2 2 3 3 3 3 3 1 3 1 6

16 3 2 1 1 3 2 3 2 3 3 2 1 3 1

17 3 3 2 3 3 3 3 1 1 1 1 1 2 5

18 2 3 1 3 1 1 4 1 1 2 1 3 3 1

19 1 1 2 1 3 3 1 2 3 1 3 1 1 4

20 2 2 3 1 3 2 4 1 1 1 3 1 1 2

21 3 3 3 1 3 1 6 2 2 1 3 1 2 5

22 1 2 1 2 3 1 4 2 3 3 3 2 2 3

23 2 1 1 2 3 2 1 1 3 3 1 1 1 5

24 1 1 2 2 2 1 3 3 2 3 3 3 3 5

25 2 3 2 1 2 3 5 3 1 3 2 1 2 1

26 2 1 3 3 3 1 2 1 2 1 2 2 3 6

27 1 3 1 3 3 2 1 3 2 3 2 2 1 2

28 3 2 2 1 2 3 4 2 1 3 3 3 1 2

29 2 2 2 2 3 1 3 3 3 3 3 2 2 4

30 3 1 2 3 2 2 2 2 3 3 2 1 3 1

31 1 2 3 1 1 2 3 3 1 1 3 3 3 4

32 2 2 1 3 1 2 5 1 1 2 2 2 1 3

33 3 3 2 1 1 1 1 1 2 1 2 2 3 6

34 1 2 2 3 3 3 5 3 1 1 1 2 1 6

35 1 1 3 2 2 2 4 2 3 1 1 3 1 6

36 2 2 1 1 2 3 2 1 3 2 2 1 1 4

37 3 3 1 2 2 2 5 2 1 2 1 1 3 4

38 1 2 3 3 2 1 1 2 3 2 2 3 2 2

39 3 3 2 3 3 3 3 2 1 3 1 2 1 5

40 1 3 1 3 3 2 1 2 2 3 2 1 3 6

41 1 1 1 3 1 1 2 2 2 3 1 3 2 4

42 2 1 1 2 2 3 3 1 2 2 1 1 2 2

43 3 2 2 2 1 2 6 2 1 3 1 2 1 5

44 2 1 1 2 3 2 1 3 2 2 1 2 3 4

45 2 1 2 3 1 2 6 3 2 3 2 2 1 2

46 1 2 3 1 1 2 3 3 1 1 3 3 3 4

47 3 3 3 3 2 2 4 2 2 2 2 3 1 3

48 2 2 2 3 2 1 1 3 1 3 1 1 3 3

49 1 2 2 3 3 3 5 3 1 3 2 1 2 1

50 3 1 3 1 1 3 3 2 3 2 2 3 2 2

51 1 2 3 3 2 1 1 2 1 2 1 1 3 4

52 2 3 3 2 1 3 1 3 2 1 1 3 2 3

53 1 3 3 2 3 3 2 3 1 1 1 2 1 6

54 2 2 2 3 2 1 1 1 1 1 1 1 2 5

5

8.2. Appendix B: The survey

Choice Question 1

Suppose Option A and Option B were the only grassland projects you could choose. Which one would you choose? Please read all the features of each

option and then check the box that represents your choice. If you do not like either option A or option B, then please choose the box marked “No

grassland project” which is Option C.

Attribute Number of Bird

Species Density of

Birds

Number of

endangered

species

Amount of

wildflowers.

Use of

prescribed

fire.

Distance to

restored area

Annual

cost to

your

household

I would

Choose

Option A

20 different

species

5 individuals

per acre

3 endangered

or threatened

species

60% covered in

wildflowers

No prescribed

burning

50 miles

$100

A

Option B

10 different

species

10 individuals

per acre

0 endangered

or threatened

species

40% covered in

wildflowers

Prescribed

burning once

every year

10 miles

$70

B

Option C

No Restoration Project

No cost C

5

9. References

Adamowicz, W., Boxall, P., Williams, M., & Louviere, J. (1998). Stated preference approaches

for measuring passive use values: Choice experiments and contingent valuation. American

Journal of Agricultural Economics, 80(1), 64-75.

Adamowicz, W., Swait, J., Boxall, P., Louviere, J., & Williams, M. (1997). Perceptions versus

objective measures of environmental quality in combined revealed and stated preference

models of environmental valuation. Journal of Environmental Economics and Management,

32(1), 65-84.

Adger, W. N., Brown, K., Cervigni, R., & Moran, D. (1995). Total economic value of forests in

mexico. Ambio, , 286-296.

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision

Processes, 50(2), 179-211.

Ajzen, I., Brown, T. C., & Rosenthal, L. H. (1996). Information bias in contingent valuation:

Effects of personal relevance, quality of information, and motivational orientation. Journal

of Environmental Economics and Management, 30(1), 43-57.

Ando, A., Camm, J., Polasky, S., & Solow, A. (1998). Species distributions, land values, and

efficient conservation. Science, 279(5359), 2126.

Ando, A. W., & Shah, P. (2010). Demand-side factors in optimal land conservation choice.

Resource and Energy Economics, 32(2), 203-221.

Askins, R. A., Askins, R., & Zickefoose, J. (2002). Restoring north america's birds: Lessons

from landscape ecology Yale Univ Pr.

Baarsma, B. E. (2003). The valuation of the IJmeer nature reserve using conjoint analysis.

Environmental and Resource Economics, 25(3), 343-356.

Bowker, J. M., & Stoll, J. R. (1988). Use of dichotomous choice nonmarket methods to value the

whooping crane resource. American Journal of Agricultural Economics, 70(2), 372-381.

Bowles, S. (1998). Endogenous preferences: The cultural consequences of markets and other

economic institutions. Journal of Economic Literature, 36(1), 75-111.

Boxall, P. C., & Adamowicz, W. L. (2002). Understanding heterogeneous preferences in random

utility models: A latent class approach. Environmental and Resource Economics, 23(4),

5

421-446.

Boxall, P. C., & Adamowicz, W. L. (2002). Understanding heterogeneous preferences in random

utility models: A latent class approach. Environmental and Resource Economics, 23(4),

421-446.

Boxall, P. C., Adamowicz, W. L., Swait, J., Williams, M., & Louviere, J. (1996). A comparison

of stated preference methods for environmental valuation. Ecological Economics, 18(3),

243-253.

Boxall, P. C., Adamowicz, W. L., Swait, J., Williams, M., & Louviere, J. (1996). A comparison

of stated preference methods for environmental valuation. Ecological Economics, 18(3),

243-253.

Boyer, T., & Polasky, S. (2004). Valuing urban wetlands: A review of non-market valuation

studies. Wetlands, 24(4), 744-755.

Brauning, D. W. (2001). Ecology and conservation of grassland birds of the western hemisphere.

The Wilson Bulletin, 113(2), 254-255.

Brennan, L. A., & Kuvlesky Jr, W. P. (2005). Invited paper: North american grassland birds: An

unfolding conservation crisis? Journal of Wildlife Management, 69(1), 1-13.

Bystrak, D. (1981). The north american breeding bird survey. Studies in Avian Biology, 6, 34-41.

Cabeza, M., Araújo, M. B., Wilson, R. J., Thomas, C. D., Cowley, M. J. R., & Moilanen, A.

(2004). Combining probabilities of occurrence with spatial reserve design. Journal of

Applied Ecology, 41(2), 252-262.

Carlsson, F., Frykblom, P., & Liljenstolpe, C. (2003). Valuing wetland attributes: An application

of choice experiments. Ecological Economics, 47(1), 95-103.

Carlsson, F., & Martinsson, P. (2001). Do hypothetical and actual marginal willingness to pay

differ in choice experiments?:: Application to the valuation of the environment. Journal of

Environmental Economics and Management, 41(2), 179-192.

Carlsson, F., Mørkbak, R., & Olsen, B. (2010). The first time is the hardest: A test of ordering

effects in choice experiments. Rapport Nr.: Working Papers in Economics 470,

Carson, R. T., Flores, N. E., & Meade, N. F. (2001). Contingent valuation: Controversies and

evidence. Environmental and Resource Economics, 19(2), 173-210.

5

Cameron T. A. & Englin J. (1997). Respondent Experience and Contingent Valuation of

Environmental Goods. Journal of Enviorenmental Economics and Management. 33, 296-

313.

Caula, S., Hvenegaard, G., & Marty, P. (2009). The influence of bird information, attitudes, and

demographics on public preferences toward urban green spaces: The case of montpellier,

france. Urban Forestry & Urban Greening, 8(2), 117-128.

Chalcraft, D. R., Wilsey, B. J., Bowles, C., & Willig, M. R. (2009). The relationship between

productivity and multiple aspects of biodiversity in six grassland communities. Biodiversity

and Conservation, 18(1), 91-104.

Christie, M., Hanley, N., Warren, J., Murphy, K., Wright, R., & Hyde, T. (2006). Valuing the

diversity of biodiversity. Ecological Economics, 58(2), 304-317.

Copeland, T. E., Sluis, W., & Howe, H. F. (2002). Fire season and dominance in an illinois

tallgrass prairie restoration. Restoration Ecology, 10(2), 315-323.

Costello, C., & Polasky, S. (2004). Dynamic reserve site selection. Resource and Energy

Economics, 26(2), 157-174.

Csuti, B., Polasky, S., Williams, P. H., Pressey, R. L., Camm, J. D., Kershaw, M., et al. (1997).

A comparison of reserve selection algorithms using data on terrestrial vertebrates in oregon.

Biological Conservation, 80(1), 83-97.

De Wit, M. P. and Blignaut, J. N. (2006). Using monetary valuation results with specific

reference to grasslands in south africa:Background information report no. 5. Retrieved

08/08, 2010, from http://www.grasslands.org.za/item.php?i_id=25

Dickson, T. L., & Wilsey, B. J. (2009). Biodiversity and tallgrass prairie decomposition: The

relative importance of species identity, evenness, richness, and micro-topography. Plant

Ecology, 201(2), 639-649.

Dillman, D. A. (2007). Mail and internet surveys: The tailored design method John Wiley &

Sons Inc.

Dissanayake, S., & Nunez, H. M. (2010). Optimal allocation of land for conservation: A general

equilibrium analysis.

Fletcher, R. J., & Koford, R. R. (2002). Habitat and landscape associations of breeding birds in

native and restored grasslands. The Journal of Wildlife Management, 66(4), 1011-1022.

5

Flynn, D. F. B., Gogol‐Prokurat, M., Nogeire, T., Molinari, N., Richers, B. T., Lin, B. B., et al.

(2009). Loss of functional diversity under land use intensification across multiple taxa.

Ecology Letters, 12(1), 22-33.

Gowdy, J. M. (2004). The revolution in welfare economics and its implications for

environmental valuation and policy. Land Economics, 80(2), 239.

Greene, W. H., & Hensher, D. A. (2003). A latent class model for discrete choice analysis:

Contrasts with mixed logit. Transportation Research Part B: Methodological, 37(8), 681-

698.

Greene, W. H., & Hensher, D. A. (2003). The mixed logit model: The state of practice and

warnings for the unwary. Transportation Research, 30B, , 133-176.

Haight, R. G., Revelle, C. S., & Snyder, S. A. (2000). An integer optimization approach to a

probabilistic reserve site selection problem. Operations Research, 48(5), 697-708.

Hanley, N., Mourato, S., & Wright, R. E. (2001). Choice modelling approaches: A superior

alternative for environmental valuatioin? Journal of Economic Surveys, 15(3), 435-462.

Hanley, N., Wright, R. E., & Koop, G. (2002). Modelling recreation demand using choice

experiments: Climbing in scotland. Environmental and Resource Economics, 22(3), 449-

466.

Hatch, D. A., Bartolome, J. W., Fehmi, J. S., & Hillyard, D. S. (1999). Effects of burning and

grazing on a coastal california grassland. Restoration Ecology, 7(4), 376-381.

Heaton, D. (2000). Conserving local habitat for declining grassland birds. Skokie, IL: Bird

Conservation Network,

Heimlich, R. E., Wiebe, K. D., Claassen, R., Gadsby, D., & House, R. M. (1998). Wetlands and

agriculture: Private interests and public benefits. Agricultural Economics Reports,

Hole, A. R. (2007). A comparison of approaches to estimating confidence intervals for

willingness to pay measures. Health Economics, 16(8), 827-840.

Hole, A. R. (2007). MIXLOGIT: Stata module to fit mixed logit models by using maximum

simulated likelihood. Statistical Software Components,

Horne, P., & Petajisto, L. (2003). Preferences for alternative moose management regimes among

finnish landowners: A choice experiment approach. Land Economics, 79(4), 472.

5

Howe, H. F. (1995). Succession and fire season in experimental prairie plantings. Ecology, 76(6),

1917-1925.

Howe, H., & Brown, J. (1999). Effects of birds and rodents on synthetic tallgrass communities.

Ecology, 80(5), 1776-1781.

Illinois Department of Natural Resources. (2000). Illinois natural resources geospatial data

clearinghouse : Illinois land cover map. Retrieved 06/06, 2010, from

http://www.isgs.uiuc.edu/nsdihome/webdocs/landcover/landcover99-00.html

Illinois Department of Natural Resources. (2010). Illinois wildlife action plan. Retrieved 04/01,

2010, from http://dnr.state.il.us/orc/Wildliferesources/theplan/

Kahneman, D. (1993). Stated willingness to pay for public goods: A psychological perspective.

Psychological Science, 4(5), 310.

Kaval, P., & Roskruge, M. (2008). The value of native birds in new zealand: Results of a

waikato survey.

Kharouba, H. M., & Kerr, J. T. (2010). Just passing through: Global change and the conservation

of biodiversity in protected areas. Biological Conservation, 143(5), 1094-1101.

Kjær, T. (2005). A review of the discrete choice experiment-with emphasis on its application in

health care. Health Economics Papers, 1

Knopf, F. L. (1994). Avian assemblages on altered grasslands. Studies in Avian Biology, 15, 247-

257.

Kuhfeld, W. F. (2005). Marketing research methods in SAS experimental design, choice,

conjoint, and graphical techniques.

Kuhfeld, W. F. (2005). Marketing research methods in SAS: Experimental design, choice,

conjoint, and graphical techniques. SAS-Institute TS-722,

Lancaster, K. J. (1966). A new approach to consumer theory. The Journal of Political Economy,

74(2), 132-157.

Lehtonen, E., Kuuluvainen, J., Pouta, E., Rekola, M., & Li, C. Z. (2003). Non-market benefits of

forest conservation in southern finland. Environmental Science & Policy, 6(3), 195-204.

Loomis, J., & Ekstrand, E. (1997). Economic benefits of critical habitat for the mexican spotted

owl: A scope test using a multiple-bounded contingent valuation survey. Journal of

5

Agricultural and Resource Economics, 22, 356-366.

Loomis, J. B., & Larson, D. M. (1994). Total economic values of increasing gray whale

populations: Results from a contingent valuation survey of visitors and households. Marine

Resource Economics, 9(3), 275-286.

Louviere, J. J., Hensher, D. A., & Swait, J. D. (2000). Stated choice methods: Analysis and

applications Cambridge Univ Pr.

Martin, L. M., Moloney, K. A., & Wilsey, B. J. (2005). An assessment of grassland restoration

success using species diversity components. Journal of Applied Ecology, 42(2), 327-336.

Martin, L. M., & Wilsey, B. J. (2006). Assessing grassland restoration success: Relative roles of

seed additions and native ungulate activities. Journal of Applied Ecology, 43(6), 1098-1109.

Meyerhoff, J., Liebe, U., & Hartje, V. (2009). Benefits of biodiversity enhancement of nature-

oriented silviculture: Evidence from two choice experiments in germany. Journal of Forest

Economics, 15(1-2), 37-58.

Milon, J. W., & Scrogin, D. (2006). Latent preferences and valuation of wetland ecosystem

restoration. Ecological Economics, 56(2), 162-175.

Nature Conservancy. (2008). America's grasslands. Retrieved 05/05, 2010, from

http://dsc.discovery.com/earth/im-interview/grasslands-map-pg.html

Noss, R. F., LaRoe, E. T., & Scott, J. M. (1995). Endangered Ecosystems of the United States: A

Preliminary Assessment of Loss and Degradation,

O'Hara, S. U., & Stagl, S. (2002). Endogenous preferences and sustainable development* 1.

Journal of Socio-Economics, 31(5), 511-527.

Önal, H. (2004). First-best, second-best, and heuristic solutions in conservation reserve site

selection. Biological Conservation, 115(1), 55-62.

Önal, H., & Briers, R. A. (2005). Designing a conservation reserve network with minimal

fragmentation: A linear integer programming approach. Environmental Modeling and

Assessment, 10(3), 193-202.

Önal, H., & Yanprechaset, P. (2007). Site accessibility and prioritization of nature reserves.

Ecological Economics, 60(4), 763-773.

Peterjohn, B. G., & Sauer, J. R. (1999). Population status of north american grassland birds from

5

the north american breeding bird survey, 1966-1996. Studies in Avian Biology, 19, 27-44.

Possingham, H., Wilson, K., Andelman, S., & Vynne, C. (2010). Protected areas: Goals,

limitations, and design.

Prairie Restorations Inc. (2010). Cost estimates for prairie restorations, Inc. contracting services.

http://www.prairieresto.com/cost_estimates.shtml Last Accessed on 08/22/2011

Pressey, R. L., Cabeza, M., Watts, M. E., Cowling, R. M., & Wilson, K. A. (2007). Conservation

planning in a changing world. Trends in Ecology & Evolution, 22(11), 583-592.

Roe, B., Boyle, K. J., & Teisl, M. F. (1996). Using conjoint analysis to derive estimates of

compensating variation* 1. Journal of Environmental Economics and Management, 31(2),

145-159.

Samson, F., & Knopf, F. (1994). Prairie conservation in north america. Bioscience, 44(6), 418-

421.

Schramm, P. (1990). Prairie restoration: A twenty-five year perspective on establishment and

management. Proceedings of the Twelfth North American Prairie Conference, 5–9.

Train, K. (2003). Discrete choice methods with simulation Cambridge Univ Pr.

Vogl, R. J. (1979). Some basic principles of grassland fire management. Environmental

Management, 3(1), 51-57.

Williams, J. C., ReVelle, C. S., & Levin, S. A. (2005). Spatial attributes and reserve design

models: A review. Environmental Modeling and Assessment, 10(3), 163-181.

Wilsey, B. J., Chalcraft, D. R., Bowles, C. M., & Willig, M. R. (2005). Relationships among

indices suggest that richness is an incomplete surrogate for grassland biodiversity. Ecology,

86(5), 1178-1184.

Woodward, R. T., & Wui, Y. S. (2001). The economic value of wetland services: A meta-

analysis. Ecological Economics, 37(2), 257-270.

Zizzo, D. J. (2003). Empirical evidence on interdependent preferences: Nature or nurture?

Cambridge Journal of Economics, 27(6), 867.